Closed gabriel-turinici closed 1 year ago
Hi Gabriel, the nyse_o
dataset is indeed the one from Section 5 of Helmbold. Unfortunately, I don't remember where I got it from and I don't think columns match the footnote in the paper. Despite being a benchmark for many papers, I personally think this dataset isn't very representative of empirical validation. The other datasets are slightly better, but also often cherry picked.
If you check out newer papers there should be links to available datasets.
thank you for the reply and confirmation, indeed some info and datasets are in papers. And I agree that somehow the data is cherry picked, the general performance of this type of algorithm is very diverse.
Hi,
I'm interested to use some algorithms and the data to compare with my ongoing research and other algos from the literature. Is there any indication on what each dataset corresponds to ? For instance, is "nyse_o" the reference mentioned in Section 5 of 'On-Line Portfolio Selection Using Multiplicative Updates' ? Are stocks given in the order set in the footnote (same page) ? BTW I guess prices were translated into multiplicative factors. Any hint for the other datasets welcome.
best regards and thank you for this nice package ! Gabriel